With recent development in networks and an increase in user request for high-quality services, a method and a device for encoding/decoding voice signals of from a narrowband to a wideband or a super wideband in communication environments have been developed.
The extension of communication bands means that almost all sound signals up to music and mixed contents as well as voices are included as an encoding target.
Accordingly, an encoding/decoding method based on transform of signals is importantly used.
A restriction in bit rates and a restriction in communication bands are present in code excited linear prediction (CELP) which is mainly used in existing voice encoding/decoding, but low bit rates have provided sound quality sufficient for conversations.
However, with recent development in communication techniques, available bit rates have increased and a high-quality voice and audio encoder has been actively developed. Accordingly, a transform-based encoding/decoding technique has been used as a technique other than the CELP having a restriction in communication bands.
Therefore, a method of using the transform-based encoding/decoding technique in parallel with the CELP or as an additional layer is considered.